The Neandertal Type Site Revisited: Interdisciplinary Investigations of Skeletal Remains from the Neander Valley, Germany

Home , Neandertal (valley), Neanderthal Museum

The Neandertal type site revisited: Interdisciplinary investigations of skeletal remains from the Neander Valley, Germany

Ralf W. Schmitz*, David Serre†, Georges Bonani‡, Susanne Feine*, Felix Hillgruber*, Heike Krainitzki§, Svante Pa¨ a¨ bo†, and Fred H. Smithʈ**

*Department of Early Prehistory and Quaternary Ecology, Institute for Pre- and Protohistory, University of Tu¨bingen, Schloss Hohentu¨bingen, D-72070 Tubingen, Germany; †Max Planck Institute for Evolutionary Anthropology, Inselstrasse 22, D-04103 Leipzig, Germany; ‡Institute of Particle Physics, Swiss Federal Institute of Technology, 8093 Zu¨rich, Switzerland; §Advanced Professional School for Technical Preparation Assistants, Markstrasse 185, D-44799 Bochum, Germany; and ʈDepartment of Sociology and Anthropology, Loyola University of Chicago, Chicago, IL 60626

Communicated by Erik Trinkaus, Washington University, St. Louis, MO, August 2, 2002 (received for review July 3, 2002) The 1856 discovery of the Neandertal type specimen (Neandertal 1) In August 1856, removal of deposits from the Kleine Feld- in western Germany marked the beginning of human paleontology hofer Grotte resulted in recovery of a calotte and 15 postcranial and initiated the longest-standing debate in the discipline: the role bones. Although the detailed events of the discovery of this of Neandertals in human evolutionary history. We report excava- skeleton are not known, the circumstances that led to the tions of cave sediments that were removed from the Feldhofer recovery of the original Neandertal skeleton were chronicled by caves in 1856. These deposits have yielded over 60 human skeletal local teacher and natural historian Johann Carl Fuhlrott (4). fragments, along with a large series of Paleolithic artifacts and Fuhlrott reported that the bones were found about 2 feet below faunal material. Our analysis of this material represents the ﬁrst the surface, and that the skeleton was originally oriented with the interdisciplinary analysis of Neandertal remains incorporating ge- skull facing the cave opening. Because the clay matrix adhered netic, direct dating, and morphological dimensions simultaneously. tightly to the individual bones, some of them, including the Three of these skeletal fragments ﬁt directly on Neandertal 1, calotte, were recognized only after they had been thrown out whereas several others have distinctively Neandertal features. At of the cave and fell the 20 m to the valley floor. Fortunately, least three individuals are represented in the skeletal sample. according to Fuhlrott, Beckershoff happened to be at the locality Radiocarbon dates for Neandertal 1, from which a mtDNA se- as the bones were initially noticed and told the workers to be on quence was determined in 1997, and a second individual indicate the lookout for other bones. Fuhlrott noted that the bones were an age of Ϸ40,000 yr for both. mtDNA analysis on the same second initially thought to be those of a cave bear, which had been found individual yields a sequence that clusters with other published in other local caves, and that this misidentification was respon- sible for the bones being collected (4). At Pieper’s invitation, Neandertal sequences. Fuhlrott visited the area at the end of August 1856 and identified the remains of Neandertal 1 as human. Only then, perhaps 2 hirteen kilometers east of Du¨sseldorf lies a nondescript weeks after the specimen was found, was he able to record the Tvalley cut through deposits of high-grade Devonian lime- circumstances of the discovery. Fuhlrott also stated that because stone by the Du¨ssel river. Before the 1850s, the valley was the bones were not considered to be of any great importance at defined by walls that rose as high as 50 m above the river and the time of recovery, the workers were not particularly careful extended for slightly less than 1 km in an east–west direction. and collected only the larger easily identifiable bones. Originally known as ‘‘Gesteins’’ or ‘‘Hundsklipp,’’ the valley The Neandertal find, first reported scientifically by Fuhlrott became known by about 1850 as the Neander Valley (Neander- and Schaaffhausen in 1857 (5, 6), was the subject of a detailed tal) in honor of Joachim Neander (1650–1680), a teacher, poet, analysis by Schaaffhausen (7) and became a focal point in the and composer of hymns who often visited the area (1). Descrip- debate about human evolution during the late 19th and 20th tions of this valley from the early 19th century note that centuries (8). However, all of these discussions concern only numerous caves and rock shelters of varying sizes were located aspects of Neandertal 1’s skeletal anatomy and interpretation along both the north and south valley walls (2). One of the thereof, because no archaeological or faunal materials were smaller caves located on the south wall was known as the Kleine reported from the site. Because there were no associated finds, Feldhofer Grotte, so named because of its proximity to the and the cave was destroyed without a scientific geological nearby large farm of Feldhof. This small cave (Ϸ3 m in width by analysis, the site has been considered undatable. Also, because 5 m in length by3minheight) had a very small mouth (less than no careful surveying was done, the exact location of the Kleine 1 m wide), which opened approximately 20 m above the valley Feldhofer Grotte was no longer known by 1900 (1). floor (1, 3). New Excavations in the Neander Valley Before the 1850s, Neandertal limestone was important for local construction, but by mid-decade, the demands of the Although it was generally assumed that the deposits removed from the Feldhofer caves were dumped somewhere in the area, Prussian construction industry reached the Neandertal. In 1854, attempts to discern their location were unsuccessful until 1997. Wilhelm Beckershoff and Friedrich Wilhelm Pieper founded the On the basis of careful archival research, Schmitz and his Actiengesellschaft fu¨r Marmorindustrie Neanderthal to facili- archaeological colleague, Ju¨rgen Thissen, were able to deduce tate local construction projects (1). To supply limestone for these the likely location of these deposits. In 1997, excavations by the projects, Beckershoff and Pieper focused quarrying operations Rheinisches Amt fu¨r Bodendenkmalpflege in Bonn under the on the south wall of the Neander Valley. As the limestone was quarried, the south wall and much of the north wall, as well as the caves located therein, were literally removed. As the Abbreviation: AMS, accelerator mass spectrometry. caves were encountered, the clay deposits in them had to be Data deposition: The sequence reported in this paper has been deposited in the GenBank removed to reduce contamination of the limestone during actual database (accession no. AY149291). quarrying. **To whom reprint requests should be addressed. E-mail: [email protected].

13342–13347 ͉ PNAS ͉ October 1, 2002 ͉ vol. 99 ͉ no. 20 www.pnas.org͞cgi͞doi͞10.1073͞pnas.192464099 Fig. 1. The NN 34 left zygomaticomaxillary specimen. Scale is in millimeters. Numbers designate centimeters. Fig. 2. The NN 52 inferior mandibular symphysis. Scale as in Fig. 1. direction of Schmitz and Thissen began in this area. These excavations led to the discovery that the base of the south valley 34, Fig. 1), and a right piece of temporal bone (NN 35) fit clearly wall was left intact, and cave sediments were located adjacent to on the Neandertal 1 calotte. The zygomaticomaxillary fragment this remnant. Excavation of these deposits resulted in the exhibits a number of distinctive features and would have been recovery of Paleolithic artifacts, Pleistocene faunal remains, and identified as a Neandertal even without the fit to Neandertal 1. some 24 fragments of human bone from the clay sediments on These features include an oblique zygomaticoalveolar margin, a the valley side of the remnant (9). That these must at least in part markedly enlarged maxillary sinus (even extending into the root represent remains from the Kleine Feldhofer Grotte was con- of the zygomatic arch), a columnar lateral orbital margin, firmed when a small piece of human bone (NN 13) was found to multiple zygomaticofacial foramina, marked muscle attachments fit exactly onto the lateral side of the left lateral femoral condyle for fibers of the temporalis muscle in the anterior temporal fossa, of Neandertal 1 (1). In 2000, additional excavations were con- and a parasagittal orientation of the infraorbital plate. This ducted at this location, and a much larger series of fauna, complex of features is distinctive for Neandertals, although the artifacts, and human skeletal fragments was recovered. Two individual traits are not unique to them (10–14). NN 35 is a cranial fragments from these excavations were found to fit onto posterior mastoid portion of the temporal preserving the pos- the original Neandertal 1 calotte. terior terminus of a large digastric sulcus exteriorly and a Preliminary analysis of the thousands of lithic artifacts recov- well-excavated sulcus for the sigmoid sinus internally. Both of ered from these deposits has shown that two specific Paleolithic these features also are characteristic but not specifically diag- assemblages are represented: Micoquian artifacts typical of the nostic of Neandertals. late Middle Paleolithic and Upper Paleolithic artifacts from the The other cranial fragments do not physically connect to, but Gravettian (1, 9). More systematic analysis of the artifacts is represent elements missing from, Neandertal 1. Two fragments underway, but the presence of Gravettian artifacts means that each of sphenoid (NN 19 and 56) and mandible (NN 52 and 61) not all of the human skeletal fragments necessarily represent and a piece of right occipital (NN 40) also exhibit features

Neandertals. Study of the faunal remains also is at the incipient indicative of Neandertals. The sphenoid (NN 19) exhibits the ANTHROPOLOGY stage, but the fauna indicates a Late Pleistocene age for the elongated body and sinus typical of Neandertals (15). The deposits. Also, systematic evidence of human modification in the occipital fragment exhibits a distinctive V-shaped sulcus that form of cut marks and impact fractures is evident on many bones. connects the foramen magnum rim to the inferiorly positioned occipitomastoid region typical of Neandertals (see ref. 12). NN Human Skeletal Remains 52 is an inferior portion of a mandibular symphysis with short The 1997 and 2000 excavations have resulted in the recovery of segments of the anterior and lingual symphyseal walls (Fig. 2). 62 new human skeletal pieces. Table 3 (which is published as This specimen has expansive digastric fossae, partially separated supporting information on the PNAS web site, www.pnas.org) by a crest for the genioglossus muscle that extends up the lingual lists and provides brief information on all human skeletal face. In each of these features, NN 52 is particularly similar to the fragments recovered. The new specimens are all smaller and͞or Krapina Neandertal mandibles (10). The anterior face is flat, more fragmentary than the bones recovered in 1856. This is not with no evidence of structures associated with a mental trigone surprising, because these specimens were thrown down an or even a mentum osseum, morphology consistent with a Ne- approximately 20-m rock face and subjected to breakage by andertal mandible. NN 61 is a segment of the mandibular subsequent quarrying activity while on the valley floor. Although incisure preserving an intersection of the crest toward the center fragmentary, many pieces of bone have been refitted to form of the condylar neck, another feature that tends to be charac- more complete elements (see Table 3). Although a complete teristic of Neandertals but is not unique to them (16). discussion of these specimens is not possible here, several issues The teeth include a single permanent mandibular element 4 2 are of immediate interest. First, are these specimens the remains (left I1), four permanent maxillary teeth (right P , right M , left of Neandertals? Second, do any of these elements represent M1 or M2, and left M3), and a single deciduous tooth (right dm2). additional portions of Neandertal 1? Third, are individuals other No teeth were recovered in 1856, so no dental specimen dupli- than Neandertal 1 represented? cates elements from Neandertal 1. All teeth but the left M3 (NN Six teeth and seven skull fragments have been identified, none 33) exhibit moderate-to-heavy occlusal and interproximal wear. of which duplicates preserved portions of Neandertal 1. Two In terms of size (Table 1), NN 55 (left I1) appears small cranial pieces, a left zygomatic and partial maxillary body (NN compared with other Neandertal samples. Molars are somewhat

Schmitz et al. PNAS ͉ October 1, 2002 ͉ vol. 99 ͉ no. 20 ͉ 13343 Table 1. Dental metrics European Wu¨ rm European Early Upper Krapina B(L)-L Neandertal B(L)-L Paleolithic B(L)-L NN no. Tooth B(L)-L M-D Mean Ϯ SD (n) (10) Mean Ϯ SD (n) (20) Mean Ϯ SD (n)*

16 M2 (L) 13.3 10.7 12.4 Ϯ 1.2 (17) 12.3 Ϯ 1.2 (19) 12.4 Ϯ 0.9 (29) 31 M2 (R) 13.1 10.5 12.4 Ϯ 1.2 (17) 12.3 Ϯ 1.2 (19) 12.4 Ϯ 0.9 (29) 33 M3 (L) 12.7 10.5 11.6 Ϯ 1.4 (6) 12.0 Ϯ 1.0 (16) 11.6 Ϯ 1.3 (24) 50 dm2 (R) 10.2 8.5 10.5 Ϯ 0.7 (11) 10.2 Ϯ 0.7 (13)* 10.4 Ϯ 0.6 (11) 51 P4 (R) 9.0 6.6 10.8 Ϯ 1.3 (14) 9.9 Ϯ 0.6 (20) 10.0 Ϯ 0.8 (22)

55 I1 (L) 6.2 4.2 7.7 Ϯ 0.5 (7) 7.2 Ϯ 0.3 (11) 6.3 Ϯ 0.5 (24)

B(L)-L, buccal (labial)–lingual dimension; M-D, mesio-distal dimension. *Data courtesy of E. Trinkaus. less worn, but mesiodistal diameters are certainly reduced by preserving most of the olecranon and coronoid processes, and a interproximal wear in all teeth except NN 33. Crown size and right ulnar shaft (NN 42), reconstructed from six fragments. NN morphology do not permit attribution of the teeth to any specific 60 exhibits a trochlear notch that faces anteriorly like many Late Pleistocene human group (17). Radiological analysis of the Neandertals rather than proximoanteriorly, as in most modern three permanent molars showed no unusual expansion of the human ulnae (12, 24). NN 42 lacks both articular surfaces, but pulp cavities (taurodontism), but the presence of this feature is the apparently truncated length of the shaft suggests the forelimb highly variable in Neandertals, especially in maxillary teeth (18). shortening characteristic of Neandertals (12, 25), and the rug- NN 31 (right M2), exhibits a 6.8-mm-long interproximal ‘‘tooth- osity of muscle attachments on the proximal shaft indicates the pick’’ groove mesially, a feature increasingly found in (but not specimen is adult. unique to) Neandertals (19). Although it is not certain whether the NN 1 humerus, NN 24 The only new postcranial piece that derives definitely from distal humerus, NN 42 right ulna, and NN 60 left ulna belong to Neandertal 1 is the small portion of left lateral condyle (NN 13) the same individual, these specimens demonstrate that at least mentioned above. Additionally, there is an adult right partial one additional adult Neandertal individual, in addition to Ne- ischium (NN 14), including the superior half of the ischial andertal 1, is present. However, the NN 50 deciduous molar tuberosity and inferior acetabulum. Neandertal 1 lacks the right could not derive from Neandertal 1 or from any other adult os coxae, and the similarities in size and morphology between represented in the sample. The heavy wear and resorbed roots NN 14 and the Neandertal 1 left ischium strongly suggest that indicate that this tooth was from an Ϸ11- to 14-year-old indi- both specimens represent the same individual. Several other vidual by recent human standards (26). Thus, NN 50 demon- postcranial specimens were recovered that do not duplicate strates the presence of minimally a third individual, a subadult, elements found in 1856 (see Table 3) and that could represent in the sample. portions of Neandertal 1. At least four postcranial specimens demonstrate the presence Accelerator Mass Spectrometry (AMS) Radiocarbon Dating of a minimum of one additional adult individual in this sample, AMS radiocarbon dating was conducted on samples of three because they duplicate elements known for Neandertal 1. NN 1 specimens: the right humerus of Neandertal 1, the second right is a Ϸ108-mm-long segment of a second right humerus, recon- humerus (NN 1), and a right tibia fragment (NN 4). The sample structed from four fragments. This specimen is slightly smaller from the Neandertal 1 right humerus had to be taken with care, and more gracile than the right humerus of Neandertal 1, but due to the use of preservatives to conserve the Neandertal 1 three features suggest NN 1 is a Neandertal. First, the midshaft skeleton. Investigation of the conservational history of the index (minimum shaft diameter ϭ 18.2 mm͞maximum shaft specimen reveals no indication that the specimen had been diameter ϭ 25.5) of 70.7 indicates the moderate platymeria saturated with preservatives, which is confirmed by the histo- characteristic of Neandertals but rarer for early modern humans. logical analysis of M. Schultz (University of Go¨ttingen). Thus, This value lies just below the mean of a Neandertal sample of the preservative is restricted only to the bone surface. To avoid right humeri (73.8 Ϯ 4.7, n ϭ 12) but further below the mean of surface contamination, the probe for dating was taken by coring an early Upper Paleolithic right humerus sample (79.6 Ϯ 5.1, n ϭ into the humeral compacta from a point exposed where a 19) (data courtesy of E. Trinkaus). Second, the deltoid tuberosity segment of bone was removed for the original DNA analysis (27) is markedly less pronounced than the pectoralis major tuberos- and parallel to the shaft’s long axis. No preservatives were used ity. This pattern is more characteristic of Neandertals, whereas on the recently excavated specimens (NN 1, NN 4), so that cores the reverse is more often true of early Upper Paleolithic humeri into the compacta for samples could be initiated directly from (21, 22). The pronounced pectoralis major tuberosity suggests the bone surface. that NN 1 likely represents an adult. Maximum breadth of the A few small cylinders 2.5 mm in diameter and Ϸ2–3mmin tuberosity is 10.4 mm, just above the mean for a sample of adult length were removed from each bone with a crown drill. The European Neandertals of 8.7 Ϯ 1.4 mm (n ϭ 6) and well above organic fraction referred to as collagen was extracted from the the mean for a sample of adult Early Upper Paleolithic humeri core samples by dissolving the inorganic part of the bones (5.1 Ϯ 0.9 mm, n ϭ 6) (ref. 12). Third, the deltoid tuberosity (calcium hydroxyapatite and calcium carbonate) with dilute extends broadly parallel to the shaft’s long axis rather than more hydrochloric acid (0.5 M HCl) at room temperature over several obliquely across the shaft. The former condition is characteristic hours. The bone in all samples was relatively well preserved, with of Neandertal and earlier humeri in Europe but can also be normal collagen content. After the complete dissolving of the found in other samples (23). There is also a distal right humeral inorganic part, the collagen was rinsed to pH 7 with ultra-pure metaphysis (NN 24), which might represent the same individual distilled water. After the chemical treatment, the samples were as NN 1, but no direct refitting of the two is possible at this time. dried in an oven at 60°C, then combusted to CO2 for 2 hours at The other two postcranial specimens that duplicate elements 950°C in evacuated and sealed quartz tubes together with copper in Neandertal 1 are an adult left proximal ulna (NN 60), oxide and silver powder. In the presence of hydrogen, the

13344 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.192464099 Schmitz et al. Table 2. Uncalibrated and calibrated AMS 14C ages for human performed twice independently. Both amplifications yielded prod- remains from the Neander Valley ucts that were cloned, and multiple clones were sequenced. The AMS-14C age, Calibrated possible amplification of nuclear insertion of mtDNA was excluded years before age, because primers specific to the Neandertal sequence do not yield a Lab. no. (sample) present ␦13C, ‰ years BC PCR product when used in amplifications from DNA extracted from 23 contemporary humans (27). Furthermore, different primer ETH-19660 (NN 1) 39,240 Ϯ 670 Ϫ20.0 Ϯ 1.2 40,052 Ϯ 409 pairs produced identical DNA sequences in regions where they ETH-19661 (NN 4) 40,360 Ϯ 760 Ϫ18.8 Ϯ 1.2 40,734 Ϯ 682 overlap, and it is highly unlikely that different primer pairs would ETH-20981 (Nean 1) 39,900 Ϯ 620 Ϫ19.6 Ϯ 1.1 40,394 Ϯ 512 all preferentially amplify one particular nuclear insertion of mtDNA rather than the organellar mtDNA (36). Calibration was accomplished by using the CALPAL program by O. Jo¨ ris and B. Weninger, University of Cologne, Cologne, Germany. Three different types of DNA sequences were seen (Fig. 4, which is published as supporting information on the PNAS web site). Two of these were identical to contemporary human DNA purified carbon dioxide gas was reduced to filamentous graphite sequences, whereas the third, which was seen in both amplifi- over a cobalt catalyst by using Vogel’s method (28, 29). The cations, carried two substitutions relative to the mtDNA refer- resulting graphite–cobalt mixtures were pressed into copper ence sequence (37). These two substitutions are seen in the discs to be used as targets in the ion source. mtDNA sequences of the Neandertal 1 individual (27) as well as The 14C͞12C and 13C͞12C ratios of the samples were deter- in a Neandertal individual from Vindija, Croatia (38). Six mined relative to the respective National Institute of Standards consecutive overlapping amplifications, each using one or two and Technology oxalic acid I standard values (30). Radiocarbon primers that preferentially amplify the DNA sequences deter- ages were calculated following the procedure described by mined from previous amplifications from NN 1, were used to Stuiver and Polach (31). The natural mass fractionation correc- determine a total of 357 bp of the hypervariable region I (Fig. tions of the samples were derived from the measured 13C͞12C 5, which is published as supporting information on the PNAS ratios and were normalized to ␦13C ϭϪ25‰ relative to the web site). Each fragment was cloned, and multiple clones from reference standard value. The radiocarbon ages are reported in two or more independent amplifications were sequenced. When years before present (1950), and the errors are at the 1␵ level different nucleotides were observed at any position in two (Table 2). The obtained ␦13C values of about Ϫ20‰ are in amplifications, one or more additional amplifications covering accordance with expectations (Table 2). that position were performed. The DNA sequence found to be All calibrated dates fall at Ϸ40,000 14C years before present reproducible in all clones from at least two amplifications was and are not statistically different from each other (Table 2). deemed to be the DNA sequence endogenous to the bone. Therefore, Neandertal 1 and the individual(s) represented by Although damage is frequent in DNA extracted from ancient remains, this procedure is expected to result in a maximal error NN 1 and NN 4 could well have lived at the same time. The dates rate of about 0.12% even under the unlikely scenario that each furthermore show that the Neandertal 1 specimen is comparable amplification starts from single DNA strands (39). in geological age to other European Neandertals from oxygen Because it has recently been shown that miscoding DNA isotope stage 3 (32) and demonstrate that both specimens are not lesions can be a frequent phenomenon in amplifications from from a time period where possible contact with early modern some ancient DNA extracts (39), all Neandertal sequences humans is indicated (33). previously determined by the Munich laboratory (27, 38, 40) Genetic Analysis have been reviewed. Two sets of apparently linked substitutions stand out as possible problems. These are positions 16107 and To gauge macromolecular preservation of these fossils, a small 16108, and 16111 and 16112, respectively, all of which carry C to fragment of a right tibial shaft (NN 4), which could possibly T substitutions relative to the mtDNA reference sequence (37) derive from Neandertal 1, and a fragment of a right humeral in the Neandertal 1 sequence (27). These substitutions were shaft (NN 1), which cannot represent Neandertal 1, were ana- observed together in all clones of two independent amplifica- ANTHROPOLOGY lyzed for amino acid composition and extent of amino acid tions, whereas a third amplification had three clones with none racemization. From each bone, a fragment of 10 mg was removed of these substitutions. The six remaining clones had two of the and hydrolyzed under acid conditions. The amino acids were substitutions (positions 16111 and 16112). According to the analyzed by high-performance liquid chromatography with o- strategy used (27, 39), these four substitutions are reproducible phtaldialdehyde precolumn labeling as described (34). NN 4 ͞ ͞ and were therefore deemed to represent the endogenous se- showed a D L ratio for aspartic acid of 0.21 and a D L ratio for quence. However, there are several reasons why these positions alanine of 0.123, indicating a high level of racemization. The should be regarded with caution. They are unusual in that they ratio of glycine to aspartic acid was 1.9, an unusually low value have been observed neither in Neandertal mtDNA sequences that could indicate that most of the amino acids analyzed do not determined subsequent to that of Neandertal 1 nor in mtDNA come from collagen proteins. Thus, DNA was not expected to be sequences from contemporary humans. Furthermore, they are ͞ preserved in NN 4 (35). By contrast, NN 1 had a D L ratio for clustered in an unusual pattern and result in an apparent aspartic acid and alanine of 0.077 and 0.023, respectively, acceleration of the Neandertal 1 DNA sequence relative to the whereas no D-leucine was detectable. The total amino acid other Neandertal DNA sequences determined to date. Thus, content of the bone was high, and the ratio of glycine to aspartic three possibilities exist. The C to T substitutions could be the acid was 6.9, indicating that the majority of the proteins could be endogenous DNA sequence of Neandertal 1 and thus represent collagen. Thus, the latter bone fulfills the criteria found to be an unusual pattern of substitutions. Alternatively, because two of compatible with retrieval of endogenous DNA from paleonto- the substitutions occur in all three amplifications, they may logical remains (34), whereas the former bone does not. represent a heteroplasmic state in Neandertal 1. A final possi- A sample of 0.5 g of bone was removed from NN 1, and DNA was bility is that they may represent deaminated cytosin residues extracted in a laboratory exclusively dedicated to ancient DNA resulting in nucleotide misincorporation during PCR. Until work by using methods previously described (27). An amplification, more samples from Neandertal 1 become available for further using primers matching segments of the hypervariable region I of analyses, this issue cannot be resolved. For the purposes of this the mitochondrial control region that are highly conserved among analysis, we have excluded these substitutions. However, their humans, Neandertals, chimpanzees, and other great apes was inclusion would not affect any conclusions drawn.

Schmitz et al. PNAS ͉ October 1, 2002 ͉ vol. 99 ͉ no. 20 ͉ 13345 Conclusion The presence of Middle Paleolithic tools and Late Pleistocene- age fauna as well as a direct AMS date of Ϸ40,000 14C years establishes, to our knowledge for the first time, an appropriate context for Neandertal 1. Questions concerning the antiquity of the original Neandertal specimen led many to reject any role for it in human evolution and generally impeded late 19th century progress in the study of human evolution (8, 44). Had the fauna and artifacts been recovered in 1856, the early history of human paleontology certainly would have been altered. The new skeletal specimens belonging to the Neandertal 1 individual exhibit Fig. 3. Neighbor-joining (NJ) tree rooted with 10 recent human sequences. anatomical features consistent with other Neandertals and sig- The same topology was found by using different tree estimation methods. The nificantly enhance our anatomical knowledge of the Neandertal numbers on each branch represent (Top to Bottom) the NJ bootstrap value, type specimen. All other specimens with diagnostic morphology the puzzle-tree support value, and the maximum parsimony bootstrap value. The ‘‘Neandertal 2’’ sequence derives from specimen NN 1. also appear to derive from Neandertals. Assessment of the total sample indicates that at least one other adult and one subadult are represented. The presence of other individuals in the sample The new mtDNA sequence determined for NN 1 carries 23 provides the opportunity to further investigate anatomical vari- differences from the contemporary human reference sequence ation in the sample from which Neandertal 1 is derived. (37) and 17 differences from the closest contemporary human The mtDNA sequence from a second Neandertal individual DNA sequence present in GenBank. By contrast, it carries one (NN 1), also directly dated to Ϸ40,000 14C years ago, makes the to four nucleotide differences from the three Neandertal Neandertal site the first Pleistocene locality to yield sequences mtDNA sequences published to date, which stem from Nean- from more than one human. As more Neandertal genetic data dertal 1 (27), Vindija in Croatia (38), and Mezmaiskaya cave in are identified, it may therefore become possible to study not only Russia (41). Neighbor-joining and maximum parsimony trees the morphological but also the molecular genetic variation of the were constructed by using MEGA 2.0 (42). A quartet maximum- Neandertal population at a time before there is any credible likelihood tree was constructed by using PUZZLE-TREE (43). Tree evidence of early modern humans in Europe (33). reconstructions show that the NN 1 mtDNA sequence falls This integrated biological approach, including the direct dat- together with the three previously determined Neandertal DNA ing, provides a more complete and reliable perspective on fossil sequences to the exclusion of contemporary humans (Fig. 3). material than any of these approaches carried out in isolation can The mean pair-wise difference between these four Neandertal deliver. Thus the collaboration of geneticists, morphologists, sequences is 1.7%. The mean pair-wise difference between four archaeologists, and dating specialists should be the norm in sequences of mtDNA drawn by chance from the databank was studies of late Pleistocene human remains. estimated from 10,000 replications as 1.8% (Ϯ 0.6, n ϭ 9,309) for extant humans, 12.1% (ϩ͞Ϫ2.5, n ϭ 28) for gorillas, and 9.6% R.W.S. dedicates this article to the memory of his father, who passed (ϩ͞Ϫ2.7, n ϭ 434) for chimpanzees. This observation may show away during its preparation. We are indebted to H. Koschik (Rheinisches that Neandertal mtDNA diversity in the Late Pleistocene was Amt fu¨r Bodendenkmalpflege, Bonn), F. G. Zehnder, and H.-E. Joachim comparable to that of contemporary humans and lower than that (Rheinisches Landesmuseum Bonn) for general support and permission to take samples; and to the Rheinisch-Westfa¨lische Kalkwerke (1997) of the great apes. Obviously considerably more than four Ne- and the Neanderthal Museum Foundation (2000) for excavation per- andertal sequences will be necessary to test this hypothesis and mission support. We thank N. Conard, H.-P. Uerpmann, S. Mu¨nzel, F. establish the parameters of Neandertal mtDNA diversity. Dammann, E. Schwaderer, and A. Czarnetzki (Universita¨tTu¨bingen), E. The new mtDNA sequence determined carries three differ- Trinkaus (Washington University), and M. Schultz (Universita¨tGo¨ttin- ences from the DNA sequence determined from Neandertal 1 gen) for assistance and support; and we thank H. Jensen (Universita¨t (27). Thus, it is clear that this bone stems from an individual Tu¨bingen) for photography. Also, we thank the staff of the Neanderthal different from and maternally unrelated to the Neandertal 1 Museum (Mettmann, Germany) for support. The Deutsche Stiftung individual. It is further noteworthy that it carries three differ- Denkmalschutz, the Ministerium fu¨rSta¨dtebau und Wohnen, Kultur und ences from the type specimen sequence, whereas it carries only Sport des Landes Nordrhein-Westfalen, the Verlagshaus Gruner und one difference from the Croatian Neandertal sequence. This Jahr (Hamburg), the Hillgruber family (Hamburg), Universita¨tTu¨- bingen, the Leakey Foundation, and H.-W. Burgartz (Ju¨chen) provided observation may suggest that strong geographical clustering of financial support (R.W.S.). The work of D.S. and S.P. is supported by the mtDNA sequences is not present in Neandertals. Again, how- Max Planck Society and the Bundesministerium fu¨r Bildung und For- ever, a much larger series of Neandertal mtDNA needs to be schung. The participation of F.H.S. is supported by the Alexander von studied to arrive at any conclusions about the phylogeography of Humboldt Stiftung. We are deeply grateful to all of these agencies and the Neandertal mtDNA gene pool. individuals.

1. Schmitz, R. W. & Thissen, J. (2000) Neandertal: Die Geschichte geht weiter 10. Smith, F. H. (1976) The Neandertal Remains from Krapina: A Descriptive and (Spektrum, Heidelberg). Comparative Study (Dept. Anthropol., Univ. of Tennessee, Knoxville). 2. Bongard, J. H. (1835) Wanderung zur Neandersho¨hle: Eine topographische Skizze 11. Smith, F. H. (1992) Trans. Philos. Soc. London Ser. B 337, 243–250. der Gegend von Erkrath an der Du¨ssel (Arnz, Du¨sseldorf, Germany). 12. Trinkaus, E. (1983) The Shanidar Neandertals (Academic, New York). 3. Fuhlrott, J. C. (1868) Verh. naturhist. Ver. preuss. Rheinl. 25, Corr. Bl., 13. Trinkaus, E. (1987) J. Hum. Evol. 16, 429–443. 62–70. 14. Wolpoff, M. H. (1999) Paleoanthropology (McGraw–Hill, New York), 2nd Ed. 4. Fuhlrott, J. C. (1859) Verh. naturhist. Ver. preuss. Rheinl. 16, 131–153. 15. Silipo, P., Dazzi, M., Feliciani, M., Guglielmi, G., Guidetti, G., Martini, S., 5. Fuhlrott, J. C. (1857) Verh. naturhist. Ver. preuss. Rheinl. 14, Corr. Bl., 50. Massani, D., Mori, S. & Tanfani, G. (1991) in The Circeo 1 Neandertal Skull: 6. Schaaffhausen, H. (1857) Verh. naturhist. Ver. preuss. Rheinl. 14, Corr. Bl., Studies and Documentation, eds. Piperno, M. & Scichilone, M. (Instituto 50–52. Poligrafico e Zecca Dello Stato, Rome), pp. 513–538. 7. Schaaffhausen, H. (1888) Der Neanderthaler Fund (Marcus, Bonn, Germany). 16. Quam, R. & Smith, F. H. (1998) in Neandertals and Modern Humans in Western 8. Trinkaus, E. & Shipman, P. (1992) The Neandertals (Knopf, New York). Asia, eds. Akazawa, T., Aoki, K. & Bar-Yosef, O. (Plenum, New York), pp. 9. Schmitz, R. W. & Thissen, J. (2000) in Neanderthals and Modern Humans— 405–421. Discussing the Transition, eds. Orschiedt, J. & Weniger, G.-C. (Neanderthal 17. Frayer, D. W. (1978) Evolution of the Dentition in Upper Paleolithic and Museum, Mettmann, Germany), pp. 267–273. Mesolithic Europe (Univ. Kansas Publ. Anthropol. 10, Lawrence, KS).

13346 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.192464099 Schmitz et al. 18. Skinner, M. F. & Sperber, G. H. (1992) Atlas of the Radiographs of Early Man 33. Churchill, S. E. & Smith, F. H. (2000) Yrbk. Phys. Anthropol. 43, 61–115. (Liss, New York). 34. Poinar, H. N., Ho¨ss, M., Bada, J. L. & Pa¨äbo, S. (1996) Science 272, 864–866. 19. Frayer, D. W. & Russell, M. D. (1987) Am. J. Phys. Anthropol. 74, 393–405. 35. Krings, M., Serre, D., Paunovic´,M.&Pa¨äbo, S. (2002) in The Vindija 20. Bermu´dez de Castro, J.-M. (1993) J. Hum. Evol. 24, 339–371. Neandertals: Catalog of the Skeletal Remains, eds. Rabeder, G., Paunovic´,M.& 21. Churchill, S. E. & Smith, F. H. (2000) Am. J. Phys. Anthropol. 112, 251–273. Grossschmidt, K. (Austrian Acad. Sci., Vienna), in press. 22. Hambu¨cken, A. (1993) C. R. Acad. Sci. Ser. II 317, 109–114. 36. Handt, O., Krings, M., Ward, R. H. & Pa¨äbo, S. (1996) Am. J. Hum. Genet. 59, 23. Carretero, J. M., Arsuaga, J. L. & Lorenzo, C. (1997) J. Hum. Evol. 33, 357–408. 368–376. 24. Churchill, S. E., Pearson, O. M., Grine, F. E., Trinkaus, E. & Holliday, T. W. 37. Anderson, S., Bankier, A. T., Barrell, B. G., de Bruijn, M. H., Coulson, A. R., (1996) J. Hum. Evol. 31, 213–237. Drouin, J., Eperon, I. C., Nierlich, D. P., Roe, B. A., Sanger, F., et al. (1981) 25. Holliday, T. (1997) Am. J. Phys. Anthropol. 104, 245–258. Nature (London) 290, 457–465. 26. White, T. D. & Folkens, P. A. (2000) Human Osteology (Academic, San Diego), 38. Krings, M., Capelli, C., Tschentscher, F., Geisert, H., Meyer, S., von Haeseler, 2nd Ed. A., Grossschmidt, K., Possnert, G., Paunovic´,M.&Pa¨äbo, S. (2000) Nat. Genet. 27. Krings, M., Stone, A., Schmitz, R. W., Krainitzki, H., Stoneking, M. & Pa¨äbo, 26, 144–146. S. (1997) Cell 90, 19–30. 39. Hofreiter, M., Jaenicke, V., Serre, D., von Haeseler, A. & Pa¨äbo, S. (2001) 28. Vogel, J. S., Southon, J. R., Nelson, D. E. & Brown, T. A. (1984) Nucl. Inst. Nucleic Acids Res. 29, 4793–4799. Meth. Phys. Res. B5, 289–293. 40. Krings, M., Geisert, H., Schmitz, R. W., Kraininzki, H. & Pa¨äbo, S. (1999) Proc. 29. Vogel, J. S., Southon, J. R. & Nelson, D. E. (1987) Nucl. Inst. Meth. Phys. Res. Natl. Acad. Sci. USA 96, 5581–5585. B29, 50–56. 41. Ovchinnikov, I. V., Go¨therstro¨m, A., Romanova, G. P., Kharitonov, V. M., 30. Bonani, G., Beer, J., Hofmann, H. J., Synal, H. A., Suter, M., Wo¨lfli, W., Lideń, K. & Goodwin, W. (2000) Nature (London) 404, 490–493. Pfleiderer, Ch., Kromer, B., Junghans, C. & Mu¨nnich, K. O. (1997) Nucl. Inst. 42. Kumar, S., Tamura, K. & Nei., M. (1993) MEGA: Molecular Evolutionary Meth. Phys. Res. B29, 87–90. Genetics Analysis (Penn. State Univ., University Park, PA). 31. Stuiver, M. & Polach, H. A. (1977) Radiocarbon 19, 355–363. 43. Strimmer, K. & Von Haessler, A. (1996) Mol. Biol. Evol. 13, 964–969. 32. Mellars, P. (1995) The Neanderthal Legacy (Princeton Univ. Press, Prince- 44. Spencer, F. (1984) in The Origin of Modern Humans, eds. Smith, F. H. & ton, NJ). Spencer, F. (Liss, New York), pp. 1–49. ANTHROPOLOGY

Schmitz et al. PNAS ͉ October 1, 2002 ͉ vol. 99 ͉ no. 20 ͉ 13347 Supporting information for Schmitz et al. (2002) Proc. Natl. Acad. Sci. USA, 10.1073/pnas.192464099

Table 3. Neandertal human remains (1997–2000)

No. Element Description NN 1* Humerus, right Splinter of lateral midshaft, maximum length 107.4 mm (proximal-- Not Neandertal 1 distal), maximum breadth 19.1 mm. Large segments of deltoid tuberosity and greater tubercle crest. (Conjoins with 2, 3, 47) NN 2* Humerus, right Splinter of ventral midshaft, maximum length 110.6 mm (proximal-- Not Neandertal 1 distal), maximum breadth 20.3 mm. Portions of greater tubercle crest, intertubercular groove, and lesser tubercle crest. (Conjoins with 1, 3, 47] NN 3* Humerus – right Splinter of posterior midshaft, maximum length 57.5 mm (proximal-- Not Neandertal 1 distal), maximum breadth 13.7 mm. (Conjoins with 1, 2, 47) NN 4 Tibia, right Ventral midshaft segment, maximum length 70.4 mm (proximal-- distal), maximum breadth 20.2 mm. Segment of anterior margin. [Conjoins with 23] NN 5 Tibia, right Ventral shaft with tibial tuberosity and anterior margin, maximum length 42.2 mm (proximal--distal), maximum breadth 33.2 mm. [Conjoins with 48] NN 6 Humerus, left (?) Proximal articular surface (head). Maximum dimensions 41.1 × 34 mm.) NN 7 Cervical vertebra Probable C-6. Complete except for spine and lateral transverse processes. Total breadth 51.2 mm. NN 8 Rib fragment Right (?), below angle; preserves costal groove. Maximum length 45.5 mm. [Conjoins with 12] NN 9 Rib fragment Maximum length 60.6 mm. NN 10 Rib fragment Right, near angle with costal groove. Maximum length 47.9 mm. NN 11 Rib fragment Small piece with section removed. Dimensions 11 × 15.9 mm. NN 12 Rib fragment Right (?), maximum length 30.9 mm. [Conjoins with 8] NN 13 Femur, left. Piece of lateral femoral condyle (29.8 × 12.3 mm dimensions on Neandertal 1 articular surface). Fits on Neandertal 1. NN 14† Innominate, right Adult ischium with ischial tuberosity and articular surface of acetabulum. Maximum dimensions 76.5 mm (cranio--caudal) × 48 mm. NN 15 Sacrum Left articular process for L5. Maximum dimensions 21 × 14.4 mm. NN 16 Maxillary molar, left Permanent complete M1 or M2. Moderate to heavy occlusal and interproximal wear. Mesial interproximal groove. NN 17 Cervical vertebra Right inferior articular surface of atlas. Maximum dimensions (articular facet) 20.5 × 17.1 mm. NN 18 Cervical vertebra Left articular process of C3-6. Maximum dimensions 17.3 × 15 mm. NN 19 Sphenoid Right adult sphenoid body with elongated sphenoid sinus. Maximum dimensions 34.4 (anteroposterior) × 17.9 mm. NN 20 Metacarpal II, right Proximal articular surface with large part of shaft. Adult. Maximum length 53 mm. NN 21 Capitate, left Complete, adult. Maximum dimensions 23.5 × 17.8 mm. NN 22 Cuneiform III, right Dorsal part of lateral cuneiform of adult. Maximum proximal--distal dimension 23.8 mm. NN 23 Tibia, right Anterior shaft frequent. Maximum proximal--distal dimension 49.2 mm, maximum breadth 19.9 mm. [Conjoins with 4] NN 24* Humerus, right Distal metaphysis. Epicondyles with superior olecranon fossa and No. Element Description Not Neandertal 1 medial epicondylar ridge. Maximum dimensions 85.4 mm. (proximal--distal) × 47.1 mm. NN 25 Tibia, right (?) Shaft fragment. Maximum dimensions 137.9 mm (proximal--distal) × 22 mm. NN 26 Cervical vertebra Body with left articular facets and lamina C 5-7. [Conjoins with 27] NN 27 Cervical vertebra Right articular facets and lamina C 5-7. [ Conjoins with 26] NN 28 Metacarpal II or III, left Shaft fragment lacking both articular surfaces. Maximum length 40 (?) mm. NN 29 Manual phalanx Distal row. Adult. Right ray 2 (?). Maximum length 21. 7 mm. NN 30 Manual phalanx, left Proximal row, complete except ulnar side of head. Ray 5. Maximum length 33. 6mm. NN 31 Maxillary molar, right Complete permanent M2. Moderate occlusal and distal interproximal wear. Heavy mesial interproximal wear. NN 32 Manual phalanx, left Pollex, proximal row. Maximum length 31.3 mm. NN 33 Maxillary molar, left Permanent M3. Light to moderate occlusal and mesial interproximal wear. No distal interproximal facet. Complete except lingual root. NN 34 Facial fragment Left zygomatic with adhering maxillary body, including lateral Neandertal 1 maxillary sinus. Maximum dimensions 66.2 mm. × 54.3 mm. Internal sinus pathology. Fits on Neandertal 1. NN 35 Temporal, right Fragment just posterior to mastoid process - portion of digastric Neandertal 1 groove and sigmoid sinus sulcus (interior) preserved. Maximum anterior-posterior length 40.6 mm. Fits on Neandertal 1. NN 36 Innominate Acetabular fossa with small segment of lunate articular surface rim. Maximum dimensions 29.3 × 24.9 mm. Right (?). NN 37 Lunate, left Complete, adult specimen. Maximum dimensions 16.8 × 20.1 mm. NN 38 Calcaneus (?) Right (?) with portion of tuberosity. NN 39 Rib fragment Right (?), just ventral to tubercle. Maximum length 36.4 mm. NN 40 Occipital Right portion of adult nuchal plane with foramen magnum rim. Maximum dimensions 36.5 mm × 25.1 mm. NN 41 Human (?) Proximal ulnar segment (?). Maximum dimensions 23 × 18.3 mm. NN 42‡ Ulna, right Shaft fragment, maximum length 61.2 mm. Supinator and Not Neandertal 1 interosseous crests partially preserved. [Conjoins with 43, 44(?), 45 (?), 46, 59] NN 43‡ Ulna, right Shaft fragment, maximum length 56.2 mm. Preserves base of Not Neandertal 1 coronoid process. [Conjoins with 42, 44 (?), 45 (?), 46, 59] NN 44‡ Ulna, right Small portion of articular surface. Found associated with, but not Not Neandertal 1 connected to, 43. Possibly conjoins with other right ulnar pieces. NN 45‡ Ulna, right Small angled piece of coronoid and radial articular surfaces. Found Not Neandertal 1 associated with, but not connected to, 43. Possibly conjoins with other right ulnar pieces. NN 46‡ Ulna, right Fragment of shaft (35.3 mm maximum length) with nutrient Not Neandertal 1 foramen. Conjoins with 42, 43, 44 (?), 45 (?), 59] NN 47* Humerus, right Mid-shaft splinter. Maximum dimensions 55.4 (Proximal--distal) × Not Neandertal 1 12 mm. [Conjoins with 1, 2, 3] NN 48 Tibia, right Anterior shaft fragment. Maximum length (proximal--distal) 45.7 mm. [Conjoins with 5] NN 49 Rib fragment Uncertain side. Midshaft, below angle. Maximum length 35.2 mm. NN 50§ Maxillary molar, right Deciduous dm2 crown with heavy occulusal and interproximal wear. Not Neandertal 1 Resorbed roots. NN 51 Premolar, right Permanent P4. Heavy occlusal and interproximal wear. NN 52 Mandible Basilar portion of symphysis with digastric fossa and genial crest. Maximum length (coronal plane) 37.8 mm. NN 53 Tubular bone splinter Unknown element. Maximum dimensions 42.4 mm (proximal-- No. Element Description distal) × 8.7 mm. NN 54 Not human NN 55 Mandibular incisor, left Permanent I1. Heavy occlusal and interproximal wear. NN 56 Sphenoid Segment of scaphoid fossa. Maximum dimensions 18.2 mm × 13.8 mm. NN 57 Manual phalanx. Left ray 2 (?), intermediate row. Arthritic proximal articular surface. NN 58 Metacarpal IV Proximal shaft, 32.1 mm in length. Left (?). NN 59‡ Ulnar shaft, right. Distal shaft with interosseous crest. Maximum length 82.7 mm. Not Neandertal. [Conjoins with 42, 43, 44 (?), 45 (?), 46]. Fracture callus on distal end. NN 60¶ Ulna, left Proximal end will olecranon and coranoid processes. Adult. Not Neandertal 1 Maximum proximal-distal length 47.2 mm. NN 61 Mandible Fragment of left mandible at condylar neck. Maximum dimensions 26.3 × 16.4 mm. NN 62 Not human NN 63 Not human NN 64 Rib II, left 89.6 mm long segment preserving neck, tubercle, and shaft to position where M. serratus anterior attachment would have been.

*Neandertal 1 preserves this portion of its right humerus. Specimen is slightly more gracile and smaller than Neandertal 1, but well-developed pectoralis major attachments suggest an adult. Platymeric shaft, relatively narrow deltoid tuberosity, and marked myologia pectoralis major attachment, and course of deltoid tuberosity relatively parallel to shaft long axis suggest a Neandertal. NN 24 may be the same individual but does not connect to the other pieces.

†This element is missing on Neandertal 1, and NN 14 is very similar to the preserved left ischial portion of Neandertal 1 in both size and morphology.

‡Neandertal 1 preserves a right ulna. This specimen must be from another adult individual.

§This deciduous molar is highly unlikely to be derived from an adult individual. Thus it cannot derive from Neandertal 1 and probably is not associated with other adult(s) at the site. Evidence of an adolescent individual in the sample.

¶Definite adult; cannot be from Neandertal 1, which possesses a left ulna. Relative equal projection of coronoid and olecranon processes yields a forward-facing semilunar notch, suggesting a Neandertal.